Well, that is not exactly how i would have solved the problem, but if it works for you, then good. If you are curious as to how i would solve the problem, then read on.

Use A Database!

Of course, seeing as how you have come along this far with your solution, switching now probably seems impossible. I understand that completely, but here is some code that keeps track of "threads" via DBI.pm and CGI.pm. I used the MySQL database with the following table creation syntax:

CREATE TABLE node (
  id int(10) NOT NULL auto_increment,
  name varchar(64) NOT NULL default '',
  parent int(10) default '0',
  PRIMARY KEY  (id)
);
[download]

use strict;
use warnings;

use DBI;
use Data::Dumper;
use CGI::Pretty qw(:standard *ul);

my %node;
my $dbh = DBI->connect(
    qw(DBI:vendor:database:host user pass),
    {RaiseError=>1}
);

print header(),start_html('messages'),
    h1('messages'),start_form,
    'New thread title: ', textfield('name'),br,
    'Parent thread: ', textfield('parent'),br,
    submit('go'),end_form,
;

insert_message(param('name'),param('parent')) if param('go');

display_messages();
print end_html;

$dbh->disconnect;

sub display_messages {
    my $sth = $dbh->prepare('select id,name,parent from node');
    $sth->execute;
    while (my $row = $sth->fetchrow_hashref()) {
        $node{$row->{id}} = { name => $row->{name} };
        push @{$node{$row->{parent}}->{children}}, $row->{id}
            if $row->{parent};
    }
    # uncomment this to see the datastructure
    #print pre(Dumper \%node),hr;

    print start_ul;
    display_thread($_) for sort {$a<=>$b} keys %node;
    print end_ul;
}

sub display_thread {
    my $id = shift;
    my $node = delete $node{$id};
    return unless $node;

    print li("$id: " . $node->{name});
    return unless $node->{children};

    print start_ul;
    display_thread($_) for @{$node->{children}};
    print end_ul;
}

sub insert_message {
    my ($name,$parent) = @_;
    my $sth  = $dbh->prepare('
        insert into node(name,parent) values(?,?)
    ');  
    $sth->execute($name,$parent);
}

# this sub is not used, but i'll leave it anyway ;)
sub get_last_id {
    return $dbh->selectcol_arrayref('SELECT last_insert_id()')->[0];
}
[download]

jeffa

L-LL-L--L-LL-L--L-LL-L--
-R--R-RR-R--R-RR-R--R-RR
B--B--B--B--B--B--B--B--
H---H---H---H---H---H---
(the triplet paradiddle with high-hat)

If i'm interpreting your code and comments correctly, you are expecting that

    $threads[$i] .= "\n<ul type=\"disc\">\n$link\n</ul>\n";
[download]

will insert 3 (or 4?) new lines after line $i. Whilst this may work, at least some of the time, it is TWWTDI (the wrong way to do it:).

From the Tie::File pod:

Inserting records that contain the record separator string is not supported by this module. It will probably produce a reasonable result, but what this result will be may change in a future version. Use 'splice' to insert records or to replace one record with several.

You should replace that line with

    splice @threads, $i, 0,  
        '<ul type=\"disc\">', "$link",'</ul>';
[download]

Which says 'insert the list (second line above) at position $i, whilst deleting nothing (0)'.

See perlfunc:splice for more details.

Ah! you are indeed correct!

The way i'm doing it now *does* work (it's the last \n that's optional, not the first), but i'll try as the POD suggests..

The important thing is that it's possible for the "parent" match to actually be the last line in the file (last element in the array) and so the array will have to be extended by n elements.

thanks!